room118solutions / branchbot

Save and restore the state of your Rails development and test databases as you work on different branches

git-rails-database-branch-hook

Save and restore the state of your Rails development and test databases as you work on different branches.

This allows you to make some changes to the structure of your database on a feature branch, then be able to quickly switch back to your master branch to make a hotfix or start a new feature branch with your original database structure and content.

Installation

Download the post-checkout file into the .git/hooks directory of your project. Be sure to keep the post-checkout name to match Git's expectations.

Alternatively, you can clone this repository to your computer then symlink the post-checkout file into the .git/hooks directory of each of your projects. This allows you to manage the file in one place and be able to easily pull down updates.

Support

Only PostgreSQL and MySQL databases are currently supported.

Non-Rails Usage

While this hook is geared towards Rails and depends on Ruby, it is very easy to use it in non-Ruby/Rails projects, so long as you have Ruby installed on your system.

This script will look for a config/database.yml file in your project's root directory, and expects it to look like this:

development:
  adapter: <adapter>
  username: <database username>
  password: <database password>
  database: <database name>

We currently support two adapters: mysql2 and postgresql. We don't actually rely on these gems, but instead use them to determine which database's command line tools we should use (mysqldump/mysql or pg_dump/psql). Use mysql2 if you are using a MySQL database, or postgresql if you are using a PostgreSQL database.

branchbot's Issues

git checkout without arguments from branch saves branch dump as master

I just hit this, didn't get a chance to dig in, but documenting it a bit for future self.

If you're on branch foo and run git checkout (no arguments, meant to run git checkout . to discard changes in my working directory), git leaves you on branch foo, but the hooks seem to cause branchbot to dump the state of the current database (branch) as master, which effectively just blew away your master dump, replacing it with your feature dump.

I'll have to come back and test this a bit more to confirm, but assuming that's reproducible, I'd be curious if we can catch this and do something better, either not dumping at all (you're not really changing branches), or at least not dumping to master.

Edit: Should also note the feature branch didn't have any commits yet, so it's head commit would have matched master at that point. Perhaps if there had been a commit already, that deviation in head commit would have caused it to dump the feature DB named after the feature, rather than after master.

Expose snapshot and restore commands

• Users should be able to snapshot their current branch on-demand
• Users should be able to select a prior manual snapshot to restore from
• The switch command should simply perform a snapshot and a restore, but not keyed to a specific time (same as current behavior where the automated snapshot is keyed to the branch)

Fails when branch name contains a slash

I found when switching from a branch named 'feature/description-of-feature', I'll get an error along the lines of:

Saving state of development database on 'feature/description-of-feature' branch...sh: 1: cannot create /home/peterh/project/.db_branch_dumps/database_development-feature/description-of-feature: Directory nonexistent
failed!
Failed to dump database. Halting.

Installer

install command should launch a guided installer that creates and configures the post-checkout hook for the user.